# Concept testing

Canonical page: https://www.questionpunk.com/support/concept-testing

Run AI-powered concept tests to evaluate logos, packaging, ads, and other creative assets with respondents.

Time: 15-30 minutes to set up

Level: Intermediate

Audience: Researchers, Designers, Brand managers, Product teams

## Steps

### 1. Start a concept test

Click the **Concept test** tile in the **Start a study** grid on Home to open a standalone concept-testing workspace. From inside an existing survey, you can also click **+ Add** and choose **Concept Test** from the AI category — this adds a summary card with an **Advanced edit** link into the same workspace.

### 2. Write the brief

In the **Brief** tab, define your concept: choose a concept type (**Logo**, **Packaging**, **Ad**, **Label**, **Icon**, or **Brand system**), write a description, set brand voice and audience context, and optionally add a color palette, mood references, and reference images.

### 3. Generate concepts

In the **Concepts** tab, generate variants with AI, review them in a grid, and regenerate, upload, or fine-tune any of them. Click **Edit in Design** on a variant to refine it on a 2D design canvas with an optional dieline overlay.

### 4. Choose a test method

The **Test** tab holds the key decision: how should respondents judge the concepts? Choose **Compare** (respondents see your concepts side by side and pick the stronger one, with the strongest advancing to rating), **Shelf** (respondents shop a real store shelf with your concept placed next to real competitors), or **Both** (shelf first, then the comparison). Set how many respondents to run before the strongest concepts move on. An **Advanced** section covers phase quotas and other tuning.

### 5. Field the test

In the **Respondents** tab, publish the test to get a shareable link, run a quick AI check with synthetic respondents while you wait for real ones, and track who has answered.

### 6. Review results

The **Results** tab shows winner scoring, respondent preference patterns, segment breakdowns, qualitative themes from interviews, highlight reels, and evolution timelines. Deliverables live here too — export as PDF or share a public concept report.

## Details

Concept testing automates the creative evaluation process. Instead of manually mocking up designs and running separate surveys, the concept engine generates variants with AI and guides respondents through a structured evaluation.

It works both as a standalone study and as a survey block: start it from the **Concept test** tile on Home for a dedicated workspace, or add a **Concept Test** block to an existing survey, which links out to the same workspace for advanced editing.

Six concept types are supported: **Logo**, **Packaging**, **Ad**, **Label**, **Icon**, and **Brand system**. Each type is optimized for its domain with appropriate image generation settings and evaluation criteria.

The workspace moves through five tabs: **Brief**, **Concepts**, **Test**, **Respondents**, and **Results**. The Test tab is where you decide how concepts get judged — **Compare** (respondents see your concepts side by side and pick the stronger one, with the strongest advancing to rating), **Shelf** (respondents shop a real store shelf with your concept next to real competitors), or **Both** (shelf first, then the comparison).

**Design** is the design-creation step of concept testing — a 2D design canvas with an optional dieline overlay for refining any generated variant. Open it from a variant's **Edit in Design** action.

AI image models generate concept variants from your brief. Choose between lo-fi models for quick exploration and hi-fi models for polished final renders. Multiple providers are available.

Results include winner scoring with confidence levels, segment-level breakdowns, qualitative theme extraction from interview transcripts, and auto-generated highlight reels. Deliverables and exports live alongside the report in the Results tab; share results via a public report link or export as PDF.

## Related articles

- [How AI interviews work](https://www.questionpunk.com/support/ai-how-it-works.md)
- [Synthetic testing with AI personas](https://www.questionpunk.com/support/synthetic-testing.md)
- [Question types guide](https://www.questionpunk.com/support/survey-design-question-types.md)
- [GitHub integration](https://www.questionpunk.com/support/integrations-github.md)
- [Packaging design](https://www.questionpunk.com/support/packaging-design.md)